evolucion8
Gawd
- Joined
- Mar 15, 2006
- Messages
- 917
While we all agree that the Phenom is just slighly slower in a per clock basis compared to Penryn, bear in mind that the Core i7 single threading performance advantage over the Penryn isn't because of a very enhanced architecture, is because besides of some minor tweaks, Turbo mode boost single thread performance overclocking the single core that is working with the thread. The Nehalem architecture isn't any wider than Penryn, is just smarter with Hyper Threading to maximize its execution units, the Penryn Front Ends are often underutilized in many scenarios.
In gaming performance where cache latency matters, Nehalem isn't much faster than Penryn, and Phenom II usually is faster than Penryn, loosing, matching or outperforming slighly the Nehalem, which its design is more oriented to the server front than anything else.
Back on topic, I heard that Crysis is limited by shader texturing fetches calculations, and current and past architectures are putting more effort in faw math shader performance rather than texturing fetches. If someone can enlight us. Otherwise, it can be a while before a card can come with the raw power to run Crysis maxed plus 4x FSAA at such high resolutions.
In gaming performance where cache latency matters, Nehalem isn't much faster than Penryn, and Phenom II usually is faster than Penryn, loosing, matching or outperforming slighly the Nehalem, which its design is more oriented to the server front than anything else.
Back on topic, I heard that Crysis is limited by shader texturing fetches calculations, and current and past architectures are putting more effort in faw math shader performance rather than texturing fetches. If someone can enlight us. Otherwise, it can be a while before a card can come with the raw power to run Crysis maxed plus 4x FSAA at such high resolutions.